Patrick Pichette, former chief financial officer of Google and new VC at iNovia, talks about product planning and evaluating projects inside a growing company in this McKinsey interview from 2011. It seems as relevant as it was 7 years ago.

“The challenge is in the planning. How do I feed the winners and hold back on the ones who aren’t performing the way they should? ”

We have a quarterly review process that examines every core product area and every core engineering area against three beacons.

“First, what did it do in the last 90 days and what will it do in the next 90 days. ”

Because in those 180 days, there’s a lot to deliver—for example, in the amount of code that has to ship out and the number of users and whether it’s going viral or not. We track these things continuously, but it’s worth taking a look at—in some cases weekly, in some cases monthly, but at least every 90 days, given where we are.

“The second beacon is what’s your trajectory? ”

Do the financial models and operating metrics for a couple of years out suggest a trajectory that is gaining or losing momentum? In some cases, are you going to need more capital expenditures because you’ll need more data? If you have a fantastic success, then you need more capacity—Google Instant, for example, sometimes generates answers to user queries before they’ve finished typing. That requires a lot of computing power.

“Then the third beacon is what’s your strategic positioning in the context of a fast-changing landscape? ”

If a competitor buys another company, what does that mean? Or if we ourselves decide to move on something this quarter, what does that mean for everything else that we have?

These beacons are very tactical and short term, with financial and operational metrics always running, and always viewed in the context of a shifting strategic landscape. For example, if we thought product growth would be X but now it’s three-quarters of X, we retune our resources accordingly. So if we had planned to hire a sales force of 200 in the expectation that a product would be ready to ship, we might delay hiring them for an additional 90 days to give engineering time to run through all the testing. And we have those kinds of conversations in most areas of the company every quarter. It takes about a week, a week and a half—and if we need to, we shift resources.

(Source: McKinsey, 2011)